More than 1,100 employees from leading AI companies are urging the U.S. government to back an international system for slowing advanced AI development if it starts moving faster than humans can safely track.¹ The letter, signed by staff from OpenAI, Anthropic, Google, Meta and others, reflects a growing view inside the industry that automation of AI research could sharply accelerate progress and outstrip existing safeguards.²

The central argument is not that AI progress should stop outright, but that governments may need tools to deliberately pace the frontier if AI systems begin improving AI systems too quickly. In the letter, signatories warn that “there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”² They also argue that no single lab is likely to slow down on its own because of “intense competitive pressure,” turning the problem into a classic coordination dilemma across companies and countries.³

That concern is increasingly shared by senior figures at the frontier labs themselves. According to Axios, OpenAI CEO Sam Altman has discussed the “need” to pace development with White House officials, while other industry leaders have floated their own ideas for global oversight and emergency braking mechanisms.¹ The Verge similarly framed the appeal as a call for government action on “automated AI,” with supporters arguing that the world may need time to build security measures and stronger oversight before the technology advances further.³

At the same time, the industry is still pushing wider deployment. In a post on X, OpenAI cofounder Greg Brockman said, “Science moves faster when more researchers have access to the best tools,” as he promoted broader access to frontier models for academics.Greg Brockman (gdb)⁴ That tension — between calls for caution and continued expansion — captures the debate now taking shape inside AI’s most powerful companies.